ECE 593:

SIGNAL, SPECTRA AND SIGNAL PROCESSING Classifications of Signals

1. Multichannel and Multidimensional Signals
Course Outline: Multichannel Signals – signals that are generated by multiple
• Introduction sources or multiple sensors.
• Discrete-Time Signal and System Multidimensional Signals – signals are functions of two or more
independent variables.
• Differential Equation for FIR and IIR Filters
2. According to characteristics of Time
• Convolution and Correlation Continuous-Time Signals – are defined for every value of time
• Z-Transforms and they take on values in the continuous interval.
• Transform Analysis of Systems Discrete-Time Signals – are defined only at certain specific
• Fourier Transforms value of time.
• Implementation of Discrete-Time Systems 3. According to characteristics of Amplitude
Continuous-Amplitude Signals – signals take on all possible
CHAPTER I: INTRODUCTION values on a finite or an infinite range.
Signal – any variable that carries or contains some kind of information Discrete-Amplitude Signals – signal take on values from a finite
that can be used. set of possible values, usually these values are equidistant and
Example of Signals: can be expressed as an integer multiple of the distance
1. Speech/Voice 4. Biomedical between two successive values.
2. Sound and Music Signals 4. According to signal model or Mathematical description
3. Video and Image 5. Radar/Sonar Deterministic Signals – signals that can be uniquely described
Signals by an explicit mathematical expression, a table of data, or a
– defined as any physical quantity that varies with time, space or any well-defined rules.
other independent variable/s. Random Signals – signals that either cannot be described to
Mathematical Description of Signal: any reasonable degree of accuracy by explicit mathematical
• As a function of one independent variable formulas, or such description is too complicated to be of any
o Time practical use.
 S1(t) = 5t
Analog and Digital Signals
• As a function of two independent variables Signal Notatio Characteristics
o Function of spatial coordinates n
 S(x,y) = 3x + 2xy +10y2 Analog x(t) Continuous-Time (CT) and Continuous-
Amplitude
Types of Signals: Sample xs(n) Discrete-Time (DT) and Continuous-
1. Analog Signals – signals that vary continuously with time and d Amplitude
represent variation of physical quantities. Quantiz xq(t) Continuous-time(CT) and Discrete-
2. Digital Signals – time assume discrete values and amplitudes are ed Amplitude
restricted to a finite number of levels. Digital xq(n) Discrete-time(DT) and Discrete-Amplitude
 5 5 5 5

2.5 2.5 2.5 2.5

0 5 10 15
t 0 5 10 15 n 0 15 t 0 5 10 15 n
Continuous Time Discrete Time
5 10
Quantized Signal Digital Signal 5 1. Sampling – Conversion of continuous-time signal into discrete-time
Signal Signal 5 levels levels signal obtained by taking “samples” of the continuous-time signal
at discrete-time instant.
2. Quantization – Converts discrete-time continuous-amplitude signal
into discrete-time, discrete-amplitude signal.

Signal Processing 3. Coding – each discrete-valued signal is represented by b-bit binary

sequences
Analog
Analog Analog
Signal
Signal
Processor
Signal Sampling of Analog Signal

x(n)

Analog-to- Digital Digital-to-

Analog Digital Digital Signal Digital Analog Analog
Signal Converter Signal Processor Signal Converter Signal

Advantages of Digital over Analog Signal Processing: 0 1 2 3 4 5 6 7 8 9 n

1. Digital programmable system allows flexibility in reconfiguring the
digital signal processing operations by simply changing the
program. xs (n ) =xa (n
T )

2. Guaranteed accuracy where:

3. Digital signals are easily stored on magnetic media (tape or disk) xs(n) = discrete-time signal
without deterioration or loss of signal fidelity xa(nT) = analog signal xa(t) at every T seconds
4. No drift in performance with age or temperature n = an integer sample number
5. Perfect reproducibility T = time interval between successive samples
= sampling Period or sample interval
ANALOG-TO-DIGITAL CONVERSION PROCESS Fs = sampling frequency; reciprocal of sampling period
1
Fs =
T
xa(t) xs(n) xq(n) 1001011
Sampler Quantizer Coder
Consider the analog signal
x a (t ) = A cos( 2πFt )
where:
A = amplitude
F = analog frequency
f = digital frequency
F cycles / sec ond cycle
f = = =
Fs sample / sec ond sample
 “If the highest frequency component in a signal is Fmax, then the
Example 1: signal should be sampled at the rate of at least 2Fmax for the samples to
Consider the two analog sinusoidal signals describe the signal completely”
x1 (t ) = cos 2π (10 )t
FS ≥ 2Fmax
x 2 (t ) = cos 2π (50 )t
Which is sampled at a rate of Fs = 40Hz. Determine the corresponding Nyquist Rate – minimum sampling rate.
discrete-time signals. FN = 2Fmax

Example 2:
Consider the analog signal
x a (t ) = 3 cos( 50πt ) +10 sin( 300 πt ) − cos( 100 πt )
(a) What are the analog signal frequencies?
(b) What are the Nyquist frequency for this signal?
(c) If the sampling rate is the same as Nyquist rate, what is the
discrete-time signal?
Aliasing – an effect that causes different continuous signal to become
indistinguishable when sampled.
Amplitude F2= 1/8 Hz

Time, sec
Time, sec
0

F1= -7/8 Hz Fs= 1Hz

1 2 3 4 5 6 7 8

Sampling Theorem
Example 3
Consider the analog sinusoidal signal
x a (t ) = 3sin( 100 πt )
Quantization of Continuous-Amplitude Signals
(a) Prepare a table for the analog signal and graph for Quantization – an approximation process
0 ≤ t ≤ 20 ms x q ( n ) =Q[ x ( n )]
(b) The signal xa(t) is sampled with a sampling rate of Fs = 300
samples per second. Determine the frequency of the discrete-time Where: xq(n) = sequence of quantization samples
signal. Q[ ] = quantizer operation
(c) Compute the sample value in one period of x(n). Sketch x(n) on x(n) = samples
the same diagram with xa(t). What is the period of discrete-time
signal in milliseconds? Two Methods Used in Quantization”
(d) What is the minimum sampling rate Fs in which x(n) reaches a 1. Truncation – eliminating the excess digits by discarding them.
peak value of 3? 2. Rounding – eliminating the excess digits by rounding the resulting
number.

Quantization Error/Noise [eq(n)] – error introduced in representing the

continuous-valued signal by a finite set of discrete-value level.
Quantization Levels (L) – values allowed in digital signal
Quantization Step or Resolution (Δ) – the distance between two
successive quantization levels
x max − x min
∆=
L −1
where: xmex and xmin = the maximum and minimum value of x(n)
respectively
The quantization error, eq(n), in ROUNDING is limited to
∆ ∆
− ≤ eq ( n ) ≤
2 2

Example 4
Consider the discrete-time signal
 0.9 n for n ≥ 0
x(n)
0 for n < 0 Coding of Quantized Samples
which is obtained by sampling analog exponential signal xa(t) = 0.9 t with The coding process in A/D converter assigns a unique binary number to
Fs = 1 Hz. Prepare a table for the discrete-time signal and graph for 0 ≤ each quantized levels.
n ≤ 9 and quantized (by rounding and truncation) the signal by one
2b ≥ L
decimal place. Determine also the Quantization error of the signal.
b ≥ log 2 L

Where: b = number of bits

n x(n) TRU eq(n) ROU eq(n)
NC ND Signal-to-Quantized Noise Ratio (SQNR)
0 – measures the quality of the output of the ADC, which
1 provides the ratio of the signal power to the noise power
2 (quantized error)
3
4 PS
SQNR = 10 log = 1.76 + 6.02 b in dB
5 PQ
6
7 Example 5
8
9 π
The discrete-time signal x(n) = 6.35 cos n is quantized with a
10
resolution (a) Δ=0.1 or (b) Δ=0.2. How many bits are required in ADC in
x(n) each case? Determine also the SQNR of the ADC. Use rounding in
quantizing.
1

0.8

0.6

0.4

0.2

n
0 2 4 6 8